deepsmoseli / bidirectiona-lstm-for-text-summarization- Goto Github PK

I trained your code 20k with articles , it gives me the same summary for all articles ?
what is your results ?

When I try to implement lstm_attentio.py I get an error as such,

Traceback (most recent call last):
File "lstm_Attention.py", line 39, in
en_shape=np.shape(train_data["article"][0])
NameError: name 'train_data' is not defined

help me correct it

In the function word2vecmodel, the model that is saved as word2vec throws error that it’s not UTF-8 encoded. Saving disabled.
FYI I’m running the code using Jupyter notebook. Thanks in advance. Let me know if I’m doing something wrong.

Hi,

1. Can you please provide a way to get word from vector embedding, using Word2Vec
   E.g Vector -> Word
2. Why is decoder trained on only 1 word of summary ?

Regards
Atul loona

Line 66 In the encoder_decoder function throws an error, TypeError: list indices must be integers or slices, not tuple.

Let me know if you need more details. Thanks.

Cloud you please check the bug in line 66 it returns like this : TypeError: list indices must be integers or slices, not tuple.

1. local variables 'filenames' referenced before assignment in cnn_daily_load.py
2. Also, should the location be pointing to questions folder like below: CNN_data="C:\Users\moseli\Documents\Masters of Information technology\Masters project\text mining\data\new CNN\cnn\questions"

decoder_inputs = Input(shape=(None,de_shape[1]))
While executing lstm_attention.py
I got an error regarding the shape of the decoder_inputs .
It is saying that not to give None and showing it incompatible types .
please help regarding this

.

Processed data used for training model is not found in the link. Please help me with the link for the data.